(62 days)
The uMI 550 PET/CT is a diagnostic imaging system that combines two existing imaging modalities - PET and CT. The quantitative distribution information of PET radiopharmaceuticals within the patient body measured by PET can assist healthcare providers in assessing the metabolic and physiological functions. CT provides diagnostic tomographic anatomical information as well as photon attenuation for the scanned region. The accurate registration and fusion of PET and CT images provides anatomical reference for the findings in the PET images.
This system is intended to be operated by qualified healthcare professionals to assist in the detection, localization, diagnosis, staging, restaging, treatment planning and treatment response evaluation for diseases and disorders in, but not limit to, oncology, cardiology and neurology.
The uMI 550 PET/CT system is a combined multi-slice X-Ray Computed Tomography and Positron Emission Tomography scanner. This system is intended to be operated by qualified healthcare professionals for performing diagnostic imaging examinations. The spatial alignment and precise image registration between PET and CT ensure the PET and CT images of the same region can be fused accurately for reading. PET measures the distribution of PET radiopharmaceuticals inside the human body quantitatively. CT produces the anatomical information of the same scanned region, and provides accurate localization for the findings in the PET images. The attenuation information contained in the CT images can be utilized in the PET image reconstruction to ensure quantitation accuracy.
The uMI 550 PET/CT system also includes a patient table, a workstation with associated software installed. The software is used for patient management, data management, scan control, image reconstruction and image reading. All patient images produced by the system conform to the DICOM 3.0 standard.
The uMI 550 PET/CT has been previously cleared by FDA via K182237. The modifications performed on the uMI 550 PET/CT (K182237) in this submission are due to the addition of HYPER Iterative and Auto-Planbox function. Meanwhile the sensitivity specification has been updated. HYPER Iterative allows more iterations while remains the image noise at an acceptable level by incorporating a noise control term into the objective function. HYPER Iterative can achieve high image contrast and quantification accuracy. Auto-Planbox plan the scan range by recognizing body parts on CT scout image. It locates the different body parts based on anatomy characteristic. The scan range is generated to cover the whole body parts according to protocol selection. This function will simplify scanning process, which will be convenient for user operation.
The information provided describes the acceptance criteria and study for the uMI 550 PET/CT system, focusing on the HYPER Iterative function and Auto-Planbox function, and an update to the sensitivity specification.
Here's the breakdown of the acceptance criteria and the studies that prove the device meets them:
1. Table of Acceptance Criteria and Reported Device Performance
The document highlights changes and comparisons to a predicate device (uMI 550, K182237). The acceptance criteria are largely implied by the "Same" remark for most specifications, indicating the device must meet the predicate's performance. The specific changes are detailed below:
| Feature | Acceptance Criteria (Proposed Device) | Reported Device Performance (Previous Submission / Predicate Reference) | Remark |
|---|---|---|---|
| Sensitivity | >=9cps/kBq @0cm, >=9cps/kBq @10cm | >=10cps/kBq @0cm, >=10cps/kBq @10cm | Updated specification: lower sensitivity claimed, justified by updated activity measurement factor and manageable by slightly longer scan time without affecting safety/effectiveness. |
| HYPER Iterative | Available | Not available | New function added. Performance verified through bench testing and clinical image evaluation. |
| Auto-Planbox | Available | Not available | New function added. Performance verified through bench testing. |
| Other PET Specs | Same as Predicate | (See document for full list like Spatial Resolution, NECR, etc.) | Passed non-clinical testing |
| CT Specs | Same as Predicate | (See document for full list like Image Resolution, Image Noise, etc.) | Passed non-clinical testing |
| Electrical Safety | Conforms to ANSI AAMI ES60601-1 | (Implicit conformance to standard) | Non-clinical testing conducted. |
| EMC | Conforms to IEC 60601-1-2 | (Implicit conformance to standard) | Non-clinical testing conducted. |
| Biocompatibility | Conforms to ISO 10993-5, ISO 10993-10 | (Implicit conformance to standard) | Patient contact materials tested. |
| Software V&V | Meets all software specifications | N/A (Internal V&V) | Testing results show all specifications met. |
| Cybersecurity | Conforms to guidance document | N/A (Internal implementation) | Conforms through process implementation. |
2. Sample Size Used for the Test Set and Data Provenance
-
HYPER Iterative Clinical Image Evaluation:
- Sample Size: 20 retrospectively collected clinical cases.
- Data Provenance: Retrospectively collected; country of origin is not specified but implied to be from the manufacturer's clinical sites, likely in China given the company's location.
-
Auto-Planbox Bench Testing:
- Sample Size: 16 group scout images.
- Data Provenance: Not explicitly stated whether these were clinical or simulated, but "scout images" are typical inputs for this function. Likely internal testing data.
3. Number of Experts Used to Establish the Ground Truth for the Test Set and Qualifications of those Experts
- Clinical Image Evaluation (for HYPER Iterative):
- Number of Experts: 3
- Qualifications: American board-certified nuclear medicine physicians.
4. Adjudication Method for the Test Set
- Clinical Image Evaluation (for HYPER Iterative): The document states that each image was read by the three physicians who provided an assessment. It does not explicitly mention an adjudication method for conflicting opinions (e.g., 2+1, 3+1). It describes independent assessments of image contrast and image quality.
5. If a Multi-Reader Multi-Case (MRMC) Comparative Effectiveness Study was done, If so, what was the effect size of how much human readers improve with AI vs without AI assistance
- A reader study was done for the HYPER Iterative function. This was a comparative study between images reconstructed with OSEM (without HYPER Iterative) and HYPER Iterative.
- Without AI vs. With AI Assistance: This study does not appear to be an AI-assisted interpretation study in the typical sense of human readers interacting with an AI output. Instead, it compares the output of two different reconstruction algorithms (OSEM vs. HYPER Iterative) as interpreted by human readers. The HYPER Iterative algorithm itself is an advanced reconstruction technique, which implicitly "assists" by improving image quality.
- Effect Size: The study results indicated that "HYPER Iterative has better image contrast than OSEM and the image quality is sufficient for clinical diagnosis." No quantitative effect size (e.g., AUC improvement, percentage increase in diagnostic accuracy) is provided in this summary. The assessment was qualitative using 3-point and 5-point scales.
6. If a Standalone (i.e. algorithm only without human-in-the-loop performance) was done
- For the HYPER Iterative function, bench testing was performed focusing on quantitative metrics and visual comparisons without human interpretation as the primary endpoint:
- Quantification accuracy and signal to noise ratio (SNR) using the NEMA IO phantom.
- Effectiveness for large weight patient (likely visual assessment of image quality, not human-in-the-loop diagnostic performance).
- Image contrast improvement on brain imaging (likely quantitative metrics or visual assessment without diagnostic interpretation by experts).
- For the Auto-Planbox function, performance evaluation was done through bench testing where the system's recognition of body parts was compared against manual annotation. This is a standalone technical validation.
7. The Type of Ground Truth Used
- HYPER Iterative Clinical Image Evaluation: The ground truth for comparative image quality and contrast was established by the subjective assessment of 3 American board-certified nuclear medicine physicians (expert consensus/rating).
- HYPER Iterative Bench Testing (NEMA phantom): Ground truth for quantification accuracy and SNR was established using a NEMA IO phantom, which has known physical characteristics.
- Auto-Planbox Bench Testing: Ground truth for body part recognition was established by "manual annotation" (expert annotation).
8. The Sample Size for the Training Set
- The document does not provide any information regarding the sample size of the training set for the HYPER Iterative or Auto-Planbox functions. These are advanced algorithms; their development typically involves large datasets. However, this specific 510(k) summary focuses on the validation of the modifications, not the initial development.
9. How the Ground Truth for the Training Set Was Established
- As the training set sample size is not provided, the method for establishing its ground truth is also not described in this document.
§ 892.1200 Emission computed tomography system.
(a)
Identification. An emission computed tomography system is a device intended to detect the location and distribution of gamma ray- and positron-emitting radionuclides in the body and produce cross-sectional images through computer reconstruction of the data. This generic type of device may include signal analysis and display equipment, patient and equipment supports, radionuclide anatomical markers, component parts, and accessories.(b)
Classification. Class II.